Note: Descriptions are shown in the official language in which they were submitted.
CA 02506362 2010-03-26
METHOD OF SEISMIC SOURCE MONITORING USING
MODELED SOURCE SIGNATURES WITH CALIBRATION FUNCTIONS
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates generally to the field of geophysical prospecting. More
particularly, the invention relates to the field of seismic data acquisition.
Specifically, the
invention is a method of seismic source monitoring using modeled source
signatures with
calibration functions.
2. Description of the Related Art
Marine seismic exploration typically employs a submerged seismic source towed
by a
ship and periodically activated to generate an acoustic wavefield (a "shot").
The wavefield
may be generated by a small explosive charge, an electric spark or arc, a
vibrator, or,
typically, a gun. The gun may be a water gun, vapor gun, or, most typically,
an airgun. Each
airgun contains a volume of air typically compressed to about 2000 psi (pounds
per square
inch) or more. An airgun abruptly releases its compressed air to create an air
bubble, leading
to an expanding sound wave in the water. The resulting wave front propagates
downwardly
into the earth beneath the water, reflects from subterranean earth layers, and
returns upwardly
toward the water surface. Seismic receivers, which are typically streamers of
hydrophones
that are also submerged and towed by the same or another ship, detect the
reflected wave
fronts, convert the detected wave fronts to electrical signals, and transmit
those signals to a
ship for storage and processing.
When a seismic source is triggered, it produces a complex output pressure
pulse in the
water. In an idealized situation in which the seismic source is a point
source, such as a single
airgun, and there is no sea surface, the emitted wave front is independent of
direction and
distance, except for spherical spreading. Converted to an electrical signal,
the output pulse of
an airgun would consist of a short wave train whose envelope displays an
initial short, fast,
positive rise in amplitude followed by several rapidly-decaying oscillations.
The recorded
wave train is called the signature of the seismic source.
In practice, a sea surface exists and is typically only meters away from the
seismic
source. The acoustic wave generated by the seismic source radiates by
spherical spreading in
all directions such that there is a downwardly traveling component as well as
an upwardly
1
CA 02506362 2010-03-26
traveling component. The water-air interface at the sea surface has a
reflection coefficient
typically close to a value of -1. The upwardly traveling component of the
acoustic wave is
reflected by the water surface and is reversed in polarity to become another
downgoing
component. This additional downgoing component is generally referred to as a
"ghost". The
ghost interferes with the direct wave to complicate the source signature.
Typically, a seismic source consists not of a single element, but of a
spatially-
distributed array of elements that generate direct arrivals plus the ghost
components. This is
particularly true of airguns, currently the most common form of marine seismic
source. The
spatial dimensions of an array of source elements may be comparable to the
wavelengths of
1o the acoustic waves themselves within the useful seismic frequency passband.
Thus, there is
no single source signature for an array. Rather, the source signature of an
array in the near-
field becomes a continuous function of both direction and distance. At
distances large
compared with the array dimensions, the dependence on distance in any
particular direction
becomes negligible. This region is called the far-field. It is the far-field
signature that is
useful for seismic data processing. For arrays of airguns, which typically
extend over spatial
dimensions of about 20 meters by 20 meters, the distance to the far-field is
on the order of
250 meters.
Although modern airguns produce stable wavefields in a laboratory situation,
the
wavefields produced by arrays of airguns deployed at sea are not so stable. In
a marine
environment, the wavefields of airgun arrays vary from shot to shot because of
physical
factors such as airgun drop-outs; sea surface conditions affecting the ghost;
and variations in
the array geometry, airgun depth, pressure, airgun timing, water velocity, or
sea temperature.
If these source variations could be monitored accurately, the source variation
data could be
used to significantly enhance the quality of the resultant seismic data.
Correcting for source
variations may be particularly important in situations such as four-
dimensional or time lapse
seismic, such as reservoir monitoring. In these situations, very small
differences in seismic
data sets may otherwise be swamped by the source variations.
Shot to shot variations in marine seismic sources are not often monitored,
primarily
because it is difficult to do. However, methods of seismic source monitoring
are currently in
use. A first method, the far-field method, typically employs measurement of
the vertically
traveling signature with a far-field hydrophone. The far-field method,
however, is unreliable
because the position of the sensor drifts, only a single point in the
wavefield is measured, and
it is difficult to position the sensor (hydrophone) the required distance from
the source. The
far-field method often requires moving the seismic survey vessels and
equipment to deeper
2
CA 02506362 2010-03-26
waters to make the far-field measurements. Thus, the far-field method is
difficult and
expensive to do.
A second method of seismic source monitoring, the near-field method, employs
measurement of near-field signatures, which can be analyzed to calculate the
whole
wavefield of the array. A primary example of this second method is known to
those of
ordinary skill in the art as the notional source method. A notional source
signature is a
normalized, idealized source signature that would be measured by a hydrophone
at one meter
distance from an airgun, if there were no surface reflection and no relative
motion between
the airgun bubble and the hydrophone. See, for example, the following two
publications.
The first publication is Ziolkowski, A., Parkes, G., Hatton, L., and Haugland,
T., "The
signature of an airgun array: Computation from near-field measurements
including
interactions", Geophysics, Vol. 47, No. 10 (October 1982), pp. 1413 - 1421.
The second
publication is Parkes, G., Ziolkowski, A., Hatton, L., and Haugland, T., "The
signature of an
airgun array: Computation from near-field measurements including interactions -
Practical
considerations", Geophysics, Vol. 48, No. 2 (February 1984), pp. 105 - 111.
The first publication, Ziolkowski et al. (1982), describes a theory of the
interactions
between bubbles produced by airguns in an array. Assuming that the bubbles are
small
compared with the wavelengths of seismic interest, the array of interacting
oscillating
bubbles is equivalent to a "notional" array of non-interacting oscillating
bubbles. If there are
n airguns in the array, then n independent measurements of the near-field
pressure field of the
full array may be used to determine the n notional source signatures. The
signature of the
array at any point in the water may then be calculated by linear superposition
of these n
notional source signatures. A spherical correction is also applied, in which
the notional
source signatures are scaled and time delayed relative to each other according
to distance and
direction. However, the number of near-field measurements must not be less
than the number
n of airguns in the array.
The second publication, Parkes et al. (1984), refines the solution of
Ziolkowski et al.
(1982) for the signature of an interacting array of airguns. An iterative
technique is applied
to calculate notional source signatures from the near-field measurements using
hydrophones
placed close (one meter) to each airgun. The amplitude variation effects of
the forward
motion of the hydrophones and the upward motion of the airgun bubbles with
respect to each
other are handled in a linear velocity model. However, continuous recording of
the near-field
signatures is required to recompute the wavefield if the radiation of the
airgun array changes
or becomes unstable.
3
CA 02506362 2010-03-26
The notional source method is further discussed in US Patents, 4,476.550;
4,476,553
and 4,868,794. The first of these patents is U.S. Patent No. 4,476,550,
"Determination of far
field signatures, for instance of seismic sources", filed August 25, 1981 and
issued October 9,
1984 to Ziolkowski, A.M. and Stoffa, P.L. The second patent, also issued on
October 9,
1984, is U.S. Patent No. 4,476,553, "Method of determining the signatures of
arrays of
marine seismic sources", issued to Ziolkowski, A., Hatton, L., Parkes, G., and
Haugland, T.
The third patent, issued to the same inventors as the `553 patent, is U.S.
Patent No.
4,868,794, "Method of accumulation data for use in determining the signatures
of arrays of
marine seismic sources", issued September 19, 1989.
The first patent, the Ziolkowski et al. `550 patent, discloses a method used
with towed
marine seismic streamers for ascertaining the far-field signature of an array
of airguns, each
of which is small compared with the wavelength of the highest frequency of
interest. The
airguns are fired so that interactions between the airguns are kept
negligible, by either
separation in time or separation in space. For separation in time, the airguns
are fired
sequentially so that each airgun generates all its significant radiation
before the next airgun is
fired. For separation in space, the airguns are fired more than one at a time,
but are separated
by a distance of at least one wavelength of the lowest frequency of interest.
The far-field
signature of each unit is measured by a pressure-sensitive detector close to
the airgun but in a
region where the phase spectrum of the pressure field is independent of
azimuth and range.
The far-field signature of the array is derived from the measured signatures
by summation.
The second and third patents, the Ziolkowski et al. `553 and `794 patents,
disclose a
method for determining the far-field signature of an array of n airguns. The
array is actuated
and the emitted pressure wave is measured by n hydrophones at n independent
points whose
positions with respect to the array are known. The n measurements are
processed to form n
simultaneous equations, which are then solved to produce n notional signatures
of the n
sources. Solving the simultaneous equations takes into account the
interactions between the
airguns. The signature of the entire array is then determined by superposing
the n notional
signatures.
The notional source method, however, has some intrinsic practical
difficulties. As
described in the Ziolkowski et al. (1982) publication, the number of seismic
sources (airguns)
must equal the number of independent measurements (hydrophones), to provide n
well-
determined simultaneous equations to solve. Thus, all n airguns and all n
hydrophones must
function at all times. In addition, the notional source method assumes that
the water-air
interface at the sea surface is a good planar reflector with a reflection
coefficient close to -1.
4
CA 02506362 2010-03-26
Otherwise, as described in the Ziolkowski et al. `553 and `794 patents, the
number of
unknown variables doubles to 2n, which means that the number of hydrophones
must double
to 2n.
Furthermore, the notional source method typically uses hydrophones
approximately
one meter from each airgun, as described in the Parkes et al. (1984)
publication. A
hydrophone placed near the airgun array records the primary source signature
from the airgun
plus a much smaller ghost reflection from the sea surface. Additionally, each
hydrophone
records contributions from all the surrounding airguns. Both the ghost
reflections and the
relative motion between the hydrophones and the bubbles created by the
hydrophones must
1o be accounted for in the notional source method. Thus, the notional source
method requires
precise measurements of the separation between airguns and hydrophones as well
as precise
measurements of the spacing between airguns in the array.
Thus, a need exists for a method for determining an accurate far-field seismic
source
signature for an array of seismic sources.
BRIEF SUMMARY OF THE INVENTION
The invention is a method of seismic source monitoring using modeled source
signatures with calibration functions. Physical parameters are measured for a
plurality of
seismic sources, preferably for each activation of the seismic sources.
Calibration functions
are obtained and the measured physical parameters and the calibration
functions are applied
to a model, which generates a calibrated source signature for the array of
seismic sources.
Alternatively, the measured physical parameters are applied to a model, which
generates a
modeled source signature, and then the calibration functions are applied to
the modeled
source signature to generate the calibrated source signature. Alternatively,
modeled source
signatures are generated for each seismic source and then the calibration
functions are applied
to the modeled source signatures to generate a calibrated source signature for
each seismic
source. Then the calibrated source signatures for each seismic source may be
combined,
preferably by linear superposition, to generate the calibrated source
signature for the array of
seismic sources.
The calibration functions are preferably determined by the following method.
Physical parameters are measured for an activation of a plurality of seismic
sources. The
measured physical parameters are applied to a model, which generates a modeled
source
signature for each seismic source. Source signatures are measured for the
activation of the
seismic sources, which generates a measured source signature for each seismic
source.
5
CA 02506362 2010-11-04
Differences are determined between the modeled source signatures and the
measured source
signatures for each seismic source. Then calibration functions are determined
for each seismic
source that correct for
In accordance with a first aspect of the present invention, there is provided
a method of
real-time seismic source monitoring during the performance of a seismic
survey, comprising:
obtaining calibration functions prior to the seismic survey for physical
parameters
utilized to model a source signature for an array of seismic sources from
comparison of
measured near-field source signatures and modeled near-field source
signatures;
measuring said physical parameters during activations of said array of seismic
sources;
applying said measured physical parameters to a model, thereby generating a
modeled
source signature for said activated array of seismic sources; and
applying said calibration functions to said modeled source signature, thereby
generating a
calibrated source signature for said activated array of seismic sources.
According to another aspect of the present invention, there is provided a
method of real-time
seismic source monitoring during the performance of a seismic survey,
comprising:
performing a seismic survey using an array of seismic sources;
obtaining calibration functions prior to the seismic survey for physical
parameters
utilized to model a source signature for the array of seismic sources from
comparison of
measured near-field source signatures and modeled near-field source
signatures;
measuring said physical parameters during activations of said array of seismic
sources;
applying said measured physical parameters to a model, thereby generating a
modeled
source signature for said activated array of seismic sources;
applying said calibration functions to said modeled source signature, thereby
generating a
calibrated source signature for said activated array of seismic sources;
using the calibrated source signature to generate accurate images of potential
hydrocarbon deposits in the earth's subsurface; and
using the accurate images to determine whether and where to drill wells.
6
CA 02506362 2010-11-04
BRIEF DESCRIPTION OF THE DRAWINGS
The invention and its advantages may be more easily understood by reference to
the
following detailed description and the attached drawings, in which:
FIG. 1 is a flowchart illustrating the processing steps of an embodiment of
the method of
the invention for seismic source monitoring;
FIG. 2 is a flowchart illustrating the processing steps of an embodiment of
the method of
the invention for constructing calibration functions; and
FIG. 3 is a plot showing a comparison between a modeled source signature and a
measured far-field source signature of a single airgun;
FIG. 4 is a plot showing a comparison between a modeled source signature and a
measured far-field source signature of an airgun array;
FIG. 5 is a plot showing a comparison between modeled and measured notional
source
signatures of an airgun, before correction of the modeled source signature by
calibration
functions;
FIG. 6 is a plot showing ,a comparison between modeled and measured notional
source
signatures of the airgun, after correction of the modeled source signature by
calibration
functions;
FIG. 7 is a plot showing a comparison between modeled and measured far-field
6a
CA 02506362 2010-03-26
source signatures of the airgun, before correction of the modeled source
signature by
calibration functions;
FIG. 8 is a plot showing a comparison between modeled and measured far-field
source signatures of the airgun, after correction of the modeled source
signature by
calibration functions;
FIG. 9 is a plot showing a comparison between modeled and measured far-field
source signatures of an airgun array, before correction of the modeled source
signature by
calibration functions;
FIG. 10 is a plot showing a comparison between modeled and measured far-field
1o source signatures of the airgun array, after correction of the modeled
source signature by
calibration functions;
FIG. 11 is a plot showing representative phase calibration functions for each
gun of
the airgun array of FIG. 9;
FIG. 12 is a plot showing representative amplitude calibration functions for
each gun
of the airgun array of FIG. 9;
FIG. 13 is a plot showing a comparison between calibrated modeled and measured
far-field source signatures of the airgun array; and
FIG. 14 is a plot showing a comparison between modeled and measured far-field
source signatures of the airgun array as in FIG. 13, after correction of the
modeled source
signature by calibration functions at a different depth.
While the invention will be described in connection with its preferred
embodiments, it
will be understood that the invention is not limited to these. On the
contrary, the invention is
intended to cover all alternatives, modifications, and equivalents that may be
included within
the scope of the invention, as defined by the appended claims.
DETAILED DESCRIPTION OF THE INVENTION
The invention is a method of seismic source monitoring for an array of seismic
sources. The invention makes use of measurements of physical parameters which
affect the
wavefield generated by the array, a computer model to estimate a source
signature for the
array from the physical parameter measurements, and calibration functions to
improve the
estimate of the source signature from the model. The method of the invention
produces a
source signature which is a continuous function of distance and direction.
Thus, the term
"source signature", when utilized with reference to the method of the
invention, will
designate a three-dimensional wavefield rather than a conventional one-
dimensional source
7
CA 02506362 2010-03-26
signature.
In one embodiment of the invention, measurements are made of physical
parameters
for an activation of the array of seismic sources. These measurements of
physical parameters
are applied to a computer model that outputs an estimate of a source signature
for the array of
seismic sources. This estimate of a source signature generated by applying
physical
parameters to the model is referred to herein as a modeled source signature.
Calibration
functions are applied to the modeled source signature to generate an improved
estimate of the
source signature of the array. This improved estimate of the source signature
after
application of the calibration functions is referred to herein as a calibrated
source signature.
In another embodiment, the calibration functions are applied to the model and
included in the model, so that a calibrated source signature is generated by
applying the
physical parameters to the model.
In another embodiment, the invention uses a model that outputs a single
estimated
source signature for the entire array of seismic sources. In another
embodiment, the
invention utilizes a model that outputs a separate estimated source signature
for each of the
seismic sources in the array. The multiple estimated source signatures may
then be combined
to give a single estimated source signature for the entire array of seismic
sources.
In a further embodiment, the calibration functions used above are obtained
from a
calibration test. In this calibration test, measurements are made of physical
parameters for an
activation of the array of seismic sources. These measurements of physical
parameters are
applied to a computer model that outputs a modeled source signature. For the
same
activation of the array of seismic sources, measurements are made of a source
signature for
each seismic source, preferably by hydrophones. These source signatures are
called
measured source signatures to distinguish them from the modeled source
signatures. The
modeled source signatures and the measured source signatures are then compared
for each
seismic source, to determine the differences between the signatures.
Calibration functions are
determined that correct for the differences between the modeled source
signatures and the
measured source signatures for each seismic source.
The method of the invention will be illustrated by embodiments employing
arrays of
guns, particularly airguns, as seismic sources, although the invention is not
restricted to
airguns as seismic sources. In the case of airguns, the preferred set of
measured physical
parameters comprises individual airgun parameters, airgun array parameters,
and
environmental parameters. These parameters include, but are not limited to,
airgun depths,
airgun pressure, airgun drop-outs, full three-dimensional geometry of the
array and sub-
8
CA 02506362 2010-03-26
arrays, airgun timing, water temperature, and atmospheric pressure. Some of
these
parameters are measured during each shot, that is, each time the entire array
of airguns is
activated. Other parameters change more slowly and can be measured less often.
The
following is a discussion of the physical parameters and their measurement.
The airgun depth parameter is needed to calculate the sea surface reflection,
or ghost.
Additionally, airgun depth can be used to calculate the external pressure on
the air bubble
generated by the airgun, which has a substantial effect on seismic source
signatures. This
pressure on the air bubble is caused by the weight of the column of water
above the airgun
plus the weight of the column of air above the water. The weight of the column
of air can be
to calculated from a measurement of atmospheric pressure. The weight of the
column of water
can be calculated from a measurement of airgun depth, if the sea surface is
calm. However, if
the sea surface is irregular, then the hydrostatic pressure at the airgun
should preferably be
measured directly, rather than calculated from an airgun depth that may not be
well-defined.
The airgun depth and hydrostatic pressure parameters are typically measured
for each shot.
The atmospheric pressure parameter, if needed, is typically not measured for
each shot when
the pressure is stable. However, atmospheric pressure can change rapidly when
a pressure
front arrives.
The airgun pressure parameter means the internal airgun pressure at the time
of airgun
firing. The gun pressures are measured by pressure sensors typically located
within the
airlines supplying the airguns, but any other means known in the art are
within the scope of
the invention. Airgun pressure is typically measured for each shot.
The airgun drop-out parameter means the airgun firing failure rate, the set of
airguns
that fail to fire on each shot. Airgun drop-out can thus have a large effect
on the source
signature generated by a source array. The airguns in an array are typically
of different sizes,
so different combinations of the airguns will generate different source
signatures for the
entire source array. Airgun drop-out is typically measured for each shot.
The three-dimensional geometry parameter means the relative positions of all
the
elements of the airgun array. Since the airgun array may also comprise sub-
arrays of airguns,
the geometry includes the relative positions of airguns within sub-arrays as
well as the
3o relative positions of sub-arrays within the entire array. The array
geometry determines how
the individual air bubbles created by the airguns interact with each other.
These bubble
interactions have a significant effect on the generation of the source
signature for the entire
source array. Relative positioning of the sub-arrays can be measured by
acoustic positioning
sensors or any other positioning system known in the art which gives the
required relative
9
CA 02506362 2010-03-26
accuracy. If the array is not rigid, then airgun geometry is typically
measured for each shot.
The airgun timing parameter means the relative timing of the firing of each
airgun.
This timing not only affects the synchronization in the total signature, but
it also has a
significant impact on the interaction effects between the airgun bubbles.
Thus, airgun timing
is preferably measured to high accuracy. Airgun timing is typically measured
for each shot.
The water temperature parameter means the water temperature at the depths of
the
airguns. Water temperature affects the emitted wavefield in two ways. First,
the water
temperature affects the thermodynamics of the expanding bubble. Second, the
water
temperature affects the density of the water and hence, the velocity of sound
in the water.
1o This change in velocity, in turn, affects the time delays, which are
coupled to the effects
caused by interaction between the bubbles generated by the different airguns.
Water
temperature can be expected to be stable over a typical array size. Thus,
temperature need
not be measured at each seismic source location on the array. It may be
sufficient to measure
at only one location for the entire array. Water temperature is measured on
the timescale on
which it varies, which need not be for every shot.
In an alternative embodiment, additional parameters describing sea surface
conditions
may be measured. The sea surface is a stable and near perfect reflector for
acoustic radiation
in good weather conditions. However, as sea conditions deteriorate, waves and
swell height
begin to affect the ghost reflection part of the source signature.
In further alternative embodiments, other physical parameters may also be
measured
and employed in a model for estimating seismic source signatures. For the
typical case of
airgun arrays, these other parameters may include, but are not limited to,
airgun types, airgun
volumes, airgun port size, and airgun port open-time. These airgun parameters
may affect the
size of the air bubbles generated by the airguns. In yet further alternative
embodiments, other
physical parameters such as water velocity, and a salinity profile of the
water may be
measured. These water parameters may affect the speed of the expansion and
movement of
the airgun bubbles. These water parameters may not change much locally, but
may be very
different in another location.
The measured physical parameters are applied to a model. The invention
preferably
uses a computer model of an array of airguns to estimate the acoustic
wavefield generated by
the air bubbles from the airguns in the array. The use of such models to
estimate source
signatures from seismic sources is well known in the art.
FIG. I shows a flowchart illustrating the processing steps of an embodiment of
the
invention. At step 101, an array of seismic sources to be monitored is
activated. Typically,
CA 02506362 2010-03-26
the seismic source array is monitored over a periodic sequence of activations
(shots).
At step 102, physical parameters are measured for the array of seismic sources
activated in step 101. Some measurements are made for each activation of the
airgun array,
while others are made less often, as needed. The physical parameters and their
measurement
are described above. The physical parameters should be those which determine
the
characteristics of the oscillating air bubbles generated by the seismic
sources in the array.
The measurements are preferably made by a system of sensors in the vicinity of
the array.
At step 103, the physical parameters measured in step 102 are applied to a
model and
a modeled source signature is generated by the model. The use of seismic
source models to
1o generate modeled source signatures is well known to those of ordinary skill
in the art. In one
embodiment, a single modeled source signature is generated by the model as an
estimated
source signature for the entire array. In another embodiment, a separate
modeled source
signature is generated by the model as an estimated source signature for each
seismic source
in the source array.
At step 104, calibration functions are applied to the modeled source
signatures from
step 103, thereby generating calibrated source signatures. In one embodiment,
in which a
single modeled source signature is generated by the model, this application of
calibration
functions to the single modeled source signature generates a single calibrated
source
signature for the entire source array. In another embodiment, in which a
modeled source
signature is generated by the model for each seismic source, this application
of calibration
functions to each modeled source signature generates a calibrated source
signature for each
seismic source.
In a further embodiment, the calibration functions are incorporated into the
seismic
source model discussed with reference to step 103. In this embodiment, the
model outputs a
calibrated source signature (or signatures) instead of a modeled source
signature (or
signatures). In another embodiment, the step of applying the calibration
functions is separate
from the model. In this embodiment, the calibration functions are applied to
the modeled
source signature (or signatures) generated by the model to generate the
calibrated source
signature (or signatures). The invention is illustrated in the flowchart in
FIG. I with step 103
of generating model source signatures from a model separate from step 104 of
applying
calibration functions to the modeled source signatures. This separation of
model and
calibration functions is for the sake of illustrative clarity only and is not
a limitation on the
invention.
At step 105, the calibrated source signatures for each of the seismic sources
in the
11
CA 02506362 2010-03-26
array are combined, if desired, to generate a single calibrated source
signature for the entire
array. Preferably, the calibrated source signatures are linearly superposed.
Linear
superposition includes, but is not limited to, linear addition and linear time-
shifting.
In a recapitulation of a preferred implementation of the invention, the
seismic source
array is activated (step 101) and the physical parameters are measured (step
102) and applied
to the model (step 103) for all the array seismic sources together. Then the
calibration
functions are applied (step 104) to the modeled source signature for each
seismic source
individually and the individual calibrated source signatures are then combined
together (step
105) to give a source signature for the entire array. All of these steps (101-
105) may then be
repeated.
The method of the invention produces a source signature which is a three-
dimensional
wavefield, as a continuous function of distance and direction, rather than
just a one-
dimensional source signature. Thus, further seismic processing, as is well
known in the art,
can be done with a three-dimensional source signature instead of a
conventional one-
dimensional source signature. One example of this further processing is
solving a three-
dimensional deconvolution of the three-dimensional source signature, rather
than just using
the vertical-traveling far-field source signature as a one-dimensional
approximation of the
source signature.
The seismic source monitoring method of the invention uses sensor measurements
of
physical parameters applied on a shot by shot basis to a model of an airgun
array. However,
the modeled source signatures generated by the model may only be accurate to a
few percent.
Therefore, the method of the invention uses calibration functions to improve
the estimated
source signatures generated by the model. These calibration functions, applied
in step 104 of
FIG. 1, correct for the differences between source signatures obtained from
the model and
source signatures obtained from field measurements.
The calibration functions are preferably obtained for each seismic source in
the
seismic source array to be monitored. The calibration functions are also
preferably obtained
for an anticipated range of values of the physical parameters measured in step
102, above.
The calibration functions may be stored in computer memory after their
construction and
retrieved as they are needed for application to modeled source signatures. In
a preferred
embodiment, the calibration functions are constructed from measurements of the
array
seismic sources made in field tests. These measurements include recording near-
field source
signatures as well as measuring physical parameters for the activated source
array. The
measured physical parameters are modeled to generate modeled source
signatures. The
12
CA 02506362 2010-03-26
modeled source signatures are compared to the measured source signatures. Then
the
calibration functions are calculated to correct for the differences between
the modeled source
signatures and the measured source signatures. FIG. 2 shows a flowchart
illustrating the
elements of an embodiment of the method of the invention for constructing the
calibration
functions.
At step 201, an array of seismic sources is activated in a sequence of shots.
Each shot
is an activation of the entire array of seismic sources. The array of seismic
sources is
preferably as discussed in step 101 of FIG. 1, above.
At step 202, physical parameters are measured for the array of seismic sources
from
1 o step 201. The physical parameters are preferably as discussed above.
At step 203, the physical parameters measured in step 202 are applied to a
model and
a modeled source signature is generated by the model for each seismic source.
The model is
preferably as discussed with reference to step 103 of FIG. 1, above.
At step 204, source signatures are measured for each seismic source for the
seismic
source activation. In a preferred embodiment, near-field source signatures are
measured by
hydrophones positioned approximately one meter from each seismic source while
the entire
seismic source array is activated. There are other possible measurements,
which could be
used as the basis of calibration. These include mid-field and far-field
measurements, or a
combination of near-, mid-, and far-field measurements. The near-field for a
typical seismic
source array extends out to a distance on the order of a few tens of meters
(approximately 25
m). The far field begins at a distance on the order of a few hundreds of
meters
(approximately 250 m). The mid-field exists at the distances in between.
Additionally,
calibration functions could be calculated from shots in which single airguns
are fired. These
field measurements should be as accurate as possible, since the quality of the
calibration
functions for the source signatures depend upon the quality of these
calibrating
measurements.
At step 205, differences between the modeled source signatures from step 203
and the
measured source signatures from step 204 are determined. In a preferred
embodiment of the
method of the invention, a computer algorithm known to those of ordinary skill
in the art
picks the primary and bubble peaks of any source signature. Applying this
algorithm to
modeled and measured signatures of a seismic source provides the information
to compare
the source signatures and then compute the corresponding calibration
functions. In one
embodiment, the phase of the modeled and measured source signatures is
compared. A
corresponding calibration function is constructed which stretches or
compresses a modeled
13
CA 02506362 2010-03-26
source signature between each pair of bubble peaks, so that the phase of the
bubble peaks
matches in the modeled and measured source signatures. Similarly, the
amplitude of the
modeled and measured source signatures is compared. This corresponding
calibration
function applies a time dependent amplitude scaling factor, calculated at the
bubble positions
and interpolated between them.
The measured source signatures from step 204 are converted, if necessary, to a
form
compatible for comparison with the modeled source signatures from step 203. In
the
preferred embodiment of the invention, the modeled source signatures are
generated by the
model in the form of notional source signatures. Thus, in the preferred
embodiment, the
measured source signatures are converted to notional source signatures for
comparison.
However, converting to notional source signatures is not a requirement of the
invention. For
example, the conversion may include, but not be limited to, removing the ghost
and
normalizing the source signature to a reference distance between source and
sensor, just as
notional source signatures have a normalized distance between source and
sensor of one
meter.
FIG. 3 is a plot showing a comparison between a modeled source signature 301,
shown as a solid line, and a measured vertically-traveling far-field source
signature 302,
shown as a dashed line, of a single 110 cubic inch airgun deployed at 6 meters
depth. In
general, the modeled source signature does a good overall job of estimating
the source
signature. Looking at specific features, the accuracy of the model in
predicting the initial
peak amplitude 303 is better than 10%, whereas the accuracy in predicting the
phase and
amplitude of the first bubble 304 is better than 5%. However, the prediction
of the phase and
amplitude of later bubbles is less accurate. Thus, the accuracy of the modeled
source
signature slowly varies over time.
FIG. 4 is a plot showing a comparison between a modeled source signature 401,
shown as a solid line, and a measured vertically-traveling far-field source
signature 402,
shown as a dashed line, of a 2920 cubic inch airgun array deployed at 6.25
meters depth. The
array is a mixture of Sleeveguns and Bolt 1500 LL guns. Again, the modeled
source
signature does a good job of estimating the main features of the source
signature of the array.
The prediction of the peak amplitude 403 of the source signature and its
primary to bubble
ratio are accurate to better than 5%. The overall shape of the modeled source
signature is
very good. The most noticeable error is in some of the phase characteristics
of the bubble at
between 100 and 120 msec.
The model utilized in step 203 is accurate at predicting changes in the source
14
CA 02506362 2010-03-26
signatures due to variations in the physical parameters from shot to shot.
However, the
absolute error in the modeled source signatures is often greater than the shot
to shot
variations in the modeled source signatures resulting from variations in the
physical
parameters. The method of the invention improves the estimated source
signatures of the
model with calibration functions.
Referring again to FIG. 2, at step 206, calibration functions for each seismic
source
are constructed that correct for the differences determined in step 205
between the modeled
source signature from step 203 and the measured source signature from step
204. The
calibration functions are discussed further below.
At step 207, it is determined if any more seismic sources remain in the array
of
seismic sources for which calibration functions have not been constructed. If
the answer is
yes, that more seismic sources remain, then the process returns to step 206 to
construct
calibration functions for another seismic source in the array. If the answer
is no, that no more
seismic sources remain, then the process continues on to step 208.
At step 208, it is determined if any more shots (activations of the entire
array of
seismic sources) remain in the sequence of shots from step 201. If the answer
is yes, that
more shots remain, then the process returns to step 201 to activate the array
of seismic
sources again. If the answer is no, that no more shots remain, then the
process continues on
to step 209 to end.
At step 209, the process ends. Calibration functions have been determined for
each
seismic source in an array for each of a sequence of shots under a range of
physical parameter
values. In a recapitulation of a preferred method of the invention for
constructing calibration
functions for a seismic source array, the entire array is activated (step 201)
and the physical
parameters are measured (step 202) and applied to the model (step 203) for all
the seismic
sources together in the array. Then source signatures are measured (step 204),
compared to
modeled source signatures (step 205), and used to derive calibration functions
(step 206) for
each seismic source individually (step 207) in the array. Then all of these
steps (201-207) are
repeated for each activation of the entire array (208).
The notional source method relies on there being as many hydrophone signals as
there
are airguns in the array. If a single hydrophone fails, the equations are
under-determined and
insoluble, which causes a serious problem for the method. The method of the
invention relies
on a variety of sensor measurements. The method of the invention is less
sensitive to failures
in the sensors. For example, if some of the sensors for physical parameters on
a sub-array
fail, then interpolation between the other sensors can still give an excellent
result. This
CA 02506362 2010-03-26
robustness gives the method of the invention a valuable advantage over the
notional source
method in a production environment.
The calibration functions are used in the invention to correct the differences
between
the source signatures from the model and the source signatures from the
measurements
without disturbing the accuracy of the relative shot to shot variations
estimated by the model.
One embodiment is a calibration function which picks the times of the positive
bubble peaks.
This allows the small phase errors in these bubble peaks to be corrected and a
straightforward
amplitude correction function to be applied. This embodiment will be referred
to as the first
calibration function. Utilizing just this first calibration function gives
good results. The
largest errors are usually in the negative parts of the calibrated source
signature, between the
bubble pulses. In a further embodiment, these errors may be minimized by
picking the
minima between bubbles, as well as the maxima defined by the bubble peaks.
In a further embodiment, each calibration function consists of a series of
correction
functions, which are tuned for and applied to each modeled source signature.
In particular,
these correction functions are applied to notional source signatures output by
the preferred
model. These are modeled source signatures output by the model that can be
superposed to
calculate the array signature in any direction. In a yet further embodiment,
there are two
correction functions.
The first correction function is a phase correction function applied to the
modeled
source signature. This phase correction function stretches or compresses a
modeled source
signature in the time domain to correct primary periodicity or phase errors in
the modeled
source signature. For example, if the second bubble period was in error by 2%,
this
correction function would gradually introduce a 2% time stretch to correct the
period. This
phase correction function, which is slowly varying in time, retains the shot
to shot variations
of the modeled source signature.
The second correction function is an amplitude correction function applied to
the
modeled source signature. This amplitude correction function multiplies the
source signature
by a scaling function in the time domain to correct primary amplitude errors
in the modeled
source signature. For example, if the amplitude of the third bubble in the
modeled signature
was 10% too low this correction function would scale that bubble up by 10%.
Again, this
amplitude correction function, which is slowly varying in time, retains the
shot to shot
variations of the modeled source signature.
In an alternative embodiment, one or more additional correction functions may
be
added to the series of correction functions. For example, an additional
correction function
16
CA 02506362 2010-03-26
may apply several further small corrections to the details of the source
signature. These
corrections could be reshaping functions. For example, if the predicted shape
of the primary
peak is slightly wrong, then the peak can be reshaped.
In a preferred embodiment, the method of the invention relies on computing a
set of
calibration functions for each airgun and applying this set to the modeled
source signatures.
The calibration functions are preferably obtained beforehand and then simply
applied to all
modeled source signatures. The invention does not preclude having different
sets of
calibration functions for different parameter ranges and interpolating between
them, for
example, a set at 2000 psi and a set at 2500 psi. The accuracy of the method
of the invention
1o may be less accurate if the array parameters vary significantly from the
values at which the
measurements at which the calibration functions were made (for example, if the
array were
deployed at a significantly different depth). However this problem can be
overcome by
making measurements of the array over the range of variation likely to be
encountered. Thus,
an additional algorithm may be incorporated to choose or interpolate the
calibration
functions.
A number of alternative embodiments for calibration functions may be described
in
relation to the first calibration function discussed above, which aligns on
bubble peaks only.
A second calibration function embodiment aligns on the centers of the full
width at half
maximum (FWHM) of each peak. A third calibration function embodiment aligns on
peaks
and the minima between the peaks. A fourth calibration function embodiment
aligns on
peaks, minima, and also on zero crossings. A fifth calibration function
embodiment aligns on
the FWHM of peaks and troughs from the fourth embodiment. Finally, a sixth
calibration
function embodiment does a source signature division after the initial
calibration done as in
the fifth embodiment.
The calibration functions sometimes show a zigzag appearance in the amplitude
correction curves. This is caused by a systematic difference between the
amplitude error of
the positive parts of the signature and the amplitude error of the negative
parts of the
signature. For example, consider a change in physical parameter, such as
depth, which
primarily affects the bubble period. For small changes in depth, a phase
correction followed
by the amplitude correction should work well. However, a change in bubble
period which
moves a peak on the changed signature onto a trough on the reference signature
might require
a very large parameter change for the first bubble, but a progressively
smaller change for later
bubbles. The result is that the amplitude correction for the negative part of
the signature is
applied to the positive part. The presence of the zigzag effect introduces
significant error at
17
CA 02506362 2010-03-26
this point.
To correct errors introduced by the zigzag effect, a further alternative
embodiment
employs a shot-adaptive method, which relies on the fact that the errors in
the modeled
source signatures are mainly systematic. Instead of computing a calibration
function as
before at the reference point, the amplitude and phase errors would be
recorded for each peak
and trough of the signature (and possibly phase errors for zero crossings).
When monitoring
the source during performance of a seismic survey, the peaks and troughs would
be picked
from the uncalibrated modeled source signatures for each shot. Shot variant
calibration
functions would then be calculated, which enforce peak alignment for the
application of the
1o amplitude correction curves. This alternative embodiment of the method of
the invention
should also improve the results when the gradient of the calibration curves is
high. If the
later parts of the source signature are of interest, then this shot-adaptive
technique could offer
an improvement.
Another alternative embodiment of the method of the invention is to add an
additional
calibration function obtained from a measured far-field source signature. A
straightforward
final amplitude and phase correction function could be used. The illustrated
method for
constructing calibration functions uses the notional source method as a
reference. The
reference calibration is done in good weather and in controlled conditions, in
which the boat
towing the sources and sensors is moving slowly. Under these conditions, the
notional source
method can be expected to be at its best. However, under seismic survey
operational
conditions, the accuracy of the notional source method deteriorates
significantly, whereas the
calibration method should maintain its accuracy. A final calibration from far-
field
measurements can increase the accuracy of the method of the invention.
A first example, shown in FIGS. 5 - 8, illustrates the use of the method of
the
invention with a single airgun. The airgun used in this example is a 150 cubic
inch
Sleevegun. FIG. 5 is a plot showing a comparison between modeled and measured
notional
source signatures of this airgun, before correction of the modeled source
signature by
calibration functions. A solid line 501 shows a modeled notional source
signature and a
dashed line 502 shows a notional signature calculated from a measured near-
field source
signature.
FIG. 6 is a plot showing a comparison between modeled and measured notional
source signatures of this airgun, after correction of the modeled source
signature by
calibration functions. The third calibration function, as described above, is
utilized. The
maxima and minima of the source signatures in FIG. 5 were picked to construct
the phase and
18
CA 02506362 2010-03-26
amplitude calibration functions. A solid line 601 shows a calibrated notional
source signature
and a dashed line 602 shows the notional signature calculated from the
measured near-field
source signature. The match is good, with the biggest discrepancy being in the
shoulder 603
leading up to the first bubble. This discrepancy could be easily corrected by
picking
additional points in this area in the determination of the calibration
functions.
FIG. 7 is a plot showing a comparison between modeled and measured far-field
source signatures of this airgun, before correction of the modeled source
signature by the first
calibration function, as described above. A solid line 701 shows a modeled far-
field source
signature and a dashed line 702 shows a measured far-field source signature.
FIG. 8 is a plot
to showing a comparison between modeled and measured far-field source
signatures of this
airgun, after correction of the modeled source signature by the first
calibration function. A
solid line 801 shows a calibrated far-field source signature and a dashed line
802 shows the
measured far-field source signature. Again, the agreement is good after
calibration.
A second example, shown in FIGS. 9 - 12, illustrates the use of the method of
the
invention with an airgun array. The array used in this example is a 730 cubic
inch airgun
array, which is a mixture of seven 20 to 100 cubic inch Sleeveguns and a 250
cubic inch Bolt
1500 LL airgun, deployed at 6 meters depth.
FIG. 9 is a plot showing a comparison between modeled and measured far-field
source signatures of this airgun array, before correction of the modeled
source signature by
the first calibration function. A solid line 901 shows a modeled far-field
source signature and
a dashed line 902 shows a far-field source signature calculated from near-
field measurements
by way of notional source signatures.
FIG. 10 is a plot showing a comparison between modeled and measured far-field
source signatures of this airgun array, after correction of the modeled source
signatures by the
first calibration function. A solid line 1001 shows a calibrated far-field
source signature and
a dashed line 1002 shows the far-field source signature calculated from near-
field
measurements by way of notional source signatures. This match is now very
good, even
using the first calibration function. An even better match can be expected,
although for more
work, with more sophisticated calibration functions.
FIGS. 11 and 12 are plots showing the representative phase 1101 and amplitude
1201
calibration functions, respectively, used to correct the modeled source
signature 1001 in FIG.
10. The calibration functions are calculated from the notional source
signatures, one for each
airgun of the array.
A third example, shown in FIGS. 13 and 14, illustrates the robustness of the
method
19
CA 02506362 2010-03-26
of the invention. FIG. 13 is a plot showing a comparison between calibrated
modeled and
measured far-field source signatures of the airgun array, after correction of
the modeled
source signature by the first calibration function. A solid line 1301 shows a
calibrated far-
field source signature and a dashed line 1302 shows the measured far-field
source signature.
The airgun array is at a depth of 6 meters. The agreement between the modeled
and
measured far-field source signatures is excellent. The sea-floor reflection
1303 is evident,
starting at about 400 msec.
FIG. 14 is a plot showing a comparison between modeled and measured far-field
source signatures of the airgun array, after correction of the modeled source
signature by the
to first calibration function at a different depth. A solid line 1401 shows a
calibrated far-field
source signature and a dashed line 1402 shows the measured far-field source
signature. In
this case, however, the airgun array is at a depth of 6.5 meters while the
calibration functions
were calculated for a depth of 6 meters. Nonetheless, the agreement is still
very good. Thus,
the method of the invention works even in less than ideal situations, in which
the measured
parameters may differ between the calibration runs and the production runs.
The sea-floor
reflection 1403 is again evident, starting at about 460 msec.
It should be understood that the preceding is merely a detailed description of
specific
embodiments of this invention and that numerous changes, modifications, and
alternatives to
the disclosed embodiments can be made in accordance with the disclosure here
without
departing from the scope of the invention. The preceding description,
therefore, is not meant
to limit the scope of the invention. Rather, the scope of the invention is to
be determined
only by the appended claims and their equivalents.
